Biometric Fish Classification of Temperate Species Using Convolutional Neural Network with Squeeze-and-Excitation Erlend Olsvik1, Christian M. D. Trinh1, Kristian Muri Knausg˚ard2( ), Arne Wiklund1, Tonje Knutsen Sørdalen3;4, Alf Ring Kleiven3, Lei Jiao1, and Morten Goodwin1 1 Centre for Artificial Intelligence Research, University of Agder, 4879, Grimstad, Norway 2 Department of Engineering Sciences, University of Agder 3 Institute of Marine Research (IMR), His, Norway 4 Department of Natural Sciences, Centre for Coastal Research (CCR) University of Agder, Kristiansand, Norway
[email protected] Abstract. Our understanding and ability to effectively monitor and manage coastal ecosystems are severely limited by observation methods. Automatic recognition of species in natural environment is a promising tool which would revolutionize video and image analysis for a wide range of applications in marine ecology. However, classifying fish from images captured by underwater cameras is in general very challenging due to noise and illumination variations in water. Previous classification meth- ods in the literature relies on filtering the images to separate the fish from the background or sharpening the images by removing background noise. This pre-filtering process may negatively impact the classification accu- racy. In this work, we propose a Convolutional Neural Network (CNN) using the Squeeze-and-Excitation (SE) architecture for classifying im- ages of fish without pre-filtering. Different from conventional schemes, this scheme is divided into two steps. The first step is to train the fish classifier via a public data set, i.e., Fish4Knowledge, without using im- age augmentation, named as pre-training. The second step is to train the classifier based on a new data set consisting of species that we are arXiv:1904.02768v1 [cs.CV] 4 Apr 2019 interested in for classification, named as post-training.